Urban traffic management and advanced video analytics

Urban Traffic in India is steadily growing and is a pressing concern for all city planners. Smart cities are trying to deal with this burgeoning issue through the implementation of technology and video surveillance. Panellists Sushil Kumar, Dy. Director General (IoT), Telecommunication Engineering Center, Dept of Telecommunications, Ministry of Communication; Rama Shankar Pandey, Managing Director, Hella India Lighting Ltd; Vinay Mishra, Sr. Vice President (Projects), Prama India Private Limited; Ajay Gupta, Head – Sales & Strategy, SCS TECH INDIA; Shamik Joshi, Head – Products & Technology, Amnex Infotechnologies Pvt. Ltd.; Advocate Pankaj Singh, CEO, Technosys Security Systems (P) Ltd

The reports coming out of different organisations and government bodies paint a bleak scenario for the road conditions in India. According to the World Bank, India accounts for almost 10% of all road crash deaths in the world, while having only 1% of all vehicles. In addition, inordinate amount of fuel is wasted every year by vehicles stuck in traffic, increasing driver stress.

The reasons for this scenario range from human conduct to poor signage to hazardous conditions. At the panel, the consensus between the experts was that in smart cities, both technology and driver intervention will be required to rectify this problem.

Ajay Gupta, Head – Sales & Strategy, SCS Tech India: “When the smart city mission was launched, citizens were asked about the problem areas they wanted addressed and traffic was one of the top answers. Cities like Delhi, Mumbai and Bengaluru, have consistently featured in TomTom Traffic Index’s top 10 most congested cities in the world. In 2018, Mumbai was no. 1.”

Influencing positive behaviour

Human behaviour is a common reason for traffic congestion and accidents. Over speeding, lane changing, failure to stop at designated spots, all of these are related to the driver. With around 80,000 traffic police in the country, manual intervention for traffic violations is a challenging task. This is where technology can influence behaviour of the driver through alerts and through automated fines.

Rama Shankar Pandey, Managing Director, Hella India Lighting: “The Challan system is a post-mortem. We can start trusting the drivers and get into the psychology of alerting and assisting them. Do the crash research, take all the big data, come up with protocols, and then in real time start assisting our drivers though alerts. ITMS will be a boon because the driver will receive so much information.”

Automated speed alerts are coming factory-fit in most new vehicles, which sends an alert to the driver once they exceed a certain speed. Certain vehicles also come with alerts on lane departure and on distance to the vehicle ahead. These will assist and alert the driver on incorrect driving, with the aim of improving driving etiquette.

The use of PA systems to reduce driving violations was also witnessed in a test case in Bhopal, where the PA system was used to get a particular driver to stop the vehicle before the stop line. This was done by tracking him through 4 junctions with constant alerts.

There might be a change in behaviour to a certain extent, but the panellists agreed that this can only do so much. A deterrent in the form of video surveillance and tracking is required for the driver to start following the rules.

Surveillance and Monitoring

Tracking and Monitoring is becoming a vital part of the traffic ecosystems, with the deployment of smart cameras and other forms of technology. With AIS 140 being possibly implemented in all forms of vehicles, this can become a standard system to track violators. Regulations exist to prevent traffic violations, but enforcement is weak across the country. Implementation of the right monitoring technology can create a system where all forms of traffic violations are caught.

Vinay Mishra, Sr Vice President (Projects), Prama India: “Monitoring is required. If there is a traffic cop standing at the traffic signal and the drivers know that somebody is watching, they will not jump the signal. This is where video surveillance and security comes into play. With advancement in technologies like Artificial Intelligence and Machine Learning, there are many products and solutions coming in.”

A point that was discussed was the issue of implementation for traffic management. If a traffic monitoring system is setup to look at a certain number of junctions, it is unable to account for the traffic coming from the radial junctions that feed into the larger route. Most requirements are set in the form of limited number of junctions or points that are problematic to police departments. There is a need for enhanced monitoring of other junctions, so that the traffic on the main routes is better regulated.

Data

Every panellist touched upon the need for more data and better analysis. The consensus was that urban traffic cannot be rectified without improving data gathering. Even research on accidents is based on data from the police agencies of the country, to ascertain the guilty party. There is limited data on understanding the cause of the accident, in order to prevent it.

Through video monitoring, tracking systems, and possibly dashcam systems, larger amounts of data would be made available. This could be fed into Artificial Intelligence (AI) and Machine Learning (ML) structures that can identify prospective problems and optimise traffic routes.

A point was raised on the need for more collaboration between stakeholders on this aspect. In the federal structure, if someone commits a violation at the border and enters another state, the operator at the command centre will need to coordinate with different agencies and authorities to track the violator. A central database would remove this concern.

Pankaj Singh, CEO, Technosys Security Systems: “If there is a challan on a particular number, we will come to know through the common database, our personnel will come to know, our 24/7 command centre will come to know, and the vehicle will be immediately tracked. If it is going on a particular road and at 100m there is a chowki, the message is sent for the vehicle to be stopped and fined.”

But this system does not exist at the moment and cross-vertical coordination is quite weak in most cities, making the traffic management process more cumbersome.

Data integration between verticals

Every smart city project has various functional verticals, like city surveillance, intelligent traffic management systems (ITMS), smart parking systems, waste management, water management, etc. With an increase in population, the pressure on all these verticals will increase. If drivers are unable to get parking information from the smart parking server and are driving around searching, it is estimated to increase traffic congestion by 25%.

Thus, the importance of sharing all this data across verticals was discussed. In the case of connected vehicles, the data would go to an automotive server, and the smart parking data would go to a parking vertical server. For optimisation, the smart city proprietary platforms require a horizontal layer. Data from all verticals, whether it is city surveillance, water management, waste management or smart parking, would come to the common service layer where it can be shared among the verticals.

This would allow for better analysis and real time sharing. Using AI, ML and other algorithms will allow for an intelligent transport system. Implementation trials of Cellular V2X (Vehicle to Everything) are also taking place in Europe and USA, which could change the dynamic along with the roll out of 5G networks.

Sushil Kumar, Dy. Director General (IoT), Dept of Telecommunications, Ministry of Communication: “We recently adopted oneM2M, after consultations with TSDSI and other stakeholders, which is the standard for IoT platform and for common service layer. Regarding the platform, oneM2M released 2 standards. BIS has taken this standard in their document on IoT reference architecture and has given it to MoHUA, which was further referred to in their documents and sent as an advisory to all smart city SPVs. Tomorrow there may be some other standards coming from IEEE or any other body, that standard should also be inter-operable with this standard. Any smart city should be able to share data with the other smart cities. That should be the prime requirement.”

This will potentially allow the country to develop solutions for local problems using international standards. One major focus was the importance of not copy-pasting executions from abroad for domestic needs. Indian roads have a unique ecosystem and driver behaviour, which requires an indigenous approach.

Edge computing

Connected vehicles are the way of the future, be it in the form of autonomous vehicles or shared vehicles. This is where the cellular V2X will come into play. A point was raised that for autonomous vehicles, a centralised platform might introduce a latency. If data had to be uploaded and ping back, it might be too late.

Shamik Joshi, Head – Products & Technology, Amnex Infotechnologies: “It is very important that we introduce the paradigm of edge computing. We should have a micro-cloud on all junctions or an edge computing environment within the car itself. This enables the car or the junction to make decisions on the fly. I think that should be done, along with tech like 5G and cellular V2X.”

A counter point was raised that implementation at each junction and within each car, in a short timespan might not be possible. A combination of both edge computing and central servers would be required. For decision making, and in particular cases, edge computing would be the answer. But the centralised database would serve a purpose of analysis and data gathering.

In addition, focusing on infrastructure while keeping future changes and requirements in mind, might help to mitigate the needs of tomorrow. Like making white lines more retro-reflective to help autonomous vehicles identify the edge of the road. Small changes in the present that could help to reduce the need for larger investments in the future.